Placeholder elements in Powerpoint files have no size

[maciejwie]

Bug

Some shapes with shape_type=MSO_SHAPE_TYPE.PLACEHOLDER in some PPTX files may have text and has_text_frame=True but without any bounding box coordinates (left, right, top, width, height = None). This causes generate_prov() to fail with an unhandled exception.

Steps to reproduce

Sample file: placeholder.pptx

from docling.document_converter import DocumentConverter
doc_converter = DocumentConverter()
doc_converter.convert('placeholder.pptx')

Results in:

File ~/dev/docling-test/.venv/lib/python3.12/site-packages/docling/backend/mspowerpoint_backend.py:106, in MsPowerpointDocumentBackend.generate_prov(self, shape, slide_ind, text)
    [104](https://file+.vscode-resource.vscode-cdn.net/dev/docling-test/~/dev/docling-test/.venv/lib/python3.12/site-packages/docling/backend/mspowerpoint_backend.py:104) width = shape.width
    [105](https://file+.vscode-resource.vscode-cdn.net/dev/docling-test/~/dev/docling-test/.venv/lib/python3.12/site-packages/docling/backend/mspowerpoint_backend.py:105) height = shape.height
--> [106](https://file+.vscode-resource.vscode-cdn.net/dev/docling-test/~/dev/docling-test/.venv/lib/python3.12/site-packages/docling/backend/mspowerpoint_backend.py:106) shape_bbox = [left, top, left + width, top + height]
    [107](https://file+.vscode-resource.vscode-cdn.net/dev/docling-test/~/dev/docling-test/.venv/lib/python3.12/site-packages/docling/backend/mspowerpoint_backend.py:107) shape_bbox = BoundingBox.from_tuple(shape_bbox, origin=CoordOrigin.BOTTOMLEFT)
    [108](https://file+.vscode-resource.vscode-cdn.net/dev/docling-test/~/dev/docling-test/.venv/lib/python3.12/site-packages/docling/backend/mspowerpoint_backend.py:108) # prov = [{"bbox": shape_bbox, "page": parent_slide, "span": [0, len(text)]}]

TypeError: unsupported operand type(s) for +: 'NoneType' and 'NoneType'

Docling version

Latest: 2.11.0.

Python version

Python 3.12.0